Sunday, November 07, 2010

Shark-Human Interactions – Aggression?

One of the most frequent questions our customers ask is whether our Sharks are aggressive.
Our standard answer is to tell them that in our experience, they are not aggressive as in wanting to bite us but rather, that they may sometimes be assertive, meaning that some will try and barge in to get some food and that in doing so, they may display a degree of posturing.
Granted our definition may be threading a very fine line – but words are powerful and we want to steer the conservation away from the usual stereotypes and towards a more neutral mindset.

In behavioral terms, such assertiveness is called agonism, defined as any behavior related to fighting and competition,including fighting, threat, defense, flight, freezing, avoidance, and appeasement.

Very unfortunately, commercial Shark dives continue to slowly inch towards ever more close and personal encounters even with the biggest, most powerful and consequently most dangerous predatory Sharks.
In that regard, it becomes ever more important to adequately interpret the language of the animals and to adapt commercial diving procedures and one’s individual behavior accordingly, something that unfortunately appears to be missing both on the side of some operators and many clients alike.

Luckily for us, Sharks are far from being misunderstood and unpredictable and reading their state of mind is quite easy.
Equally luckily for us, they appear unendingly and uncomprehendingly tolerant and will generally gracefully overlook our breach of etiquette, meaning that people are likely to only get into trouble if they persist on displaying the most reckless, stupid and persistent of behaviors – which some individuals, being notoriously unpredictable, will undoubtedly manage to achieve!
Case in point!

The following paper is a useful compilation of the agonistic behaviors of several species of Sharks known in 2006. Yes it is long (!) but it’s an excellent reference for any Shark diver and thus, I believe, required reading. This is not the entire paper but merely excerpts that I have chosen for their immediate correlation to Shark diving.
More thoughts below.

A review of shark agonistic displays: comparison of display features and implications for shark-human interactions

: R. Aidan Martin
: Sharks are potentially dangerous animals. While the author offers his opinion on actions that may make encounters with sharks less dangerous, neither he nor Taylor & Francis accept any responsibility whatsoever for the safety of those who choose to interact with sharks, regardless of whether or not they follow the author's safety suggestions. Those who seek out interactions with sharks do so entirely at their own risk.


Agonistic displays in 23 species of sharks of six families are described and illustrated.
These displays are reviewed in terms of
ethological concepts and shark hydrodynamic models. Shark agonistic displays feature many common elements rendering them readily distinguishable from normal swimming and pseudodisplays caused by sharksucker irritation. Shark agonistic displays are most readily elicited by rapid, direct diver approach when food is absent and potential escape routes restricted. Such displays appear to be motivated by defence of self or the immediately surrounding space rather than defence of territory or resources. Costs and benefits of display versus attack in shark-shark and shark-diver contests are evaluated using payoff matrices and optimal strategies are identified. Shark-human interactions are modelled in terms of a system of nested critical approach distances. For divers faced with a displaying shark, responses which may decrease the likelihood of defensive attack are suggested. Recommendations for future work on shark agonistic behaviour are offered.


Agonism is a class of behaviour that occurs in competitive contexts and is important to the social dynamics and evolutionary fitness in animals of many taxa (Hall and Halliday 1998).
For example, individuals displaced or defeated in agonistic contests often disperse into adjacent habitats that are frequently suboptimal, thus reducing their feeding and reproductive success (Reese 1978). Agonistic behaviour is widespread among animals, including crustaceans, insects, spiders, mammals, birds, reptiles, amphibians, and fishes (see Bradbury and Vehrencamp 1998, for a recent review). Agonistic displays have been recorded in many teleost species, representing numerous families and orders (Fine et al. 1977; Colgan 1993; Guthrie and Muntz 1993; Helfman et al. 1997; Reebs 2001).

Many shark species exhibit a variety of so
cial behaviours, yet agonistic displays have rarely been reported in this group (Bres 1993).
Shark agonistic displays are best known in the grey reef shark (
Carcharhinus amblyrhynchos), which exhibits a distinctive, highly stereotyped ‘‘hunch display’’ (named for the characteristic raised snout-arched back posture, resembling a sustained shrug) in response to persistent crowding or oriented pursuit by divers or submersibles ( Johnson and Nelson 1973; Doak 1975; Stark and Anderson 1978; Nelson et al. 1986).
This display features four postural elements,
(1) a prominently raised snout,
(2) bilaterally (to_the same angle) depressed pectoral fins,
(3) arched back, and
(4) lateral flexure of the body, combined with two locomotory elements, (a) laterally exaggerated swimming, and (b) rolling and/or spiral looping ( Johnson and Nelson 1973).

The hunch display of the grey reef shark is a graded response, increasing in intensity with speed and directness of diver approach as well as degree of restriction of the shark’s escape route ( Johnson and Nelson 1973).
If a pursuing diver or submersible persists, the hunch display rapidly escalates to either a rapid withdrawal or a lightning-fast, open-mouthed slashing attack ( Johnson and Nelson 1973; Nelson et al. 1986) Thus, the hunch display of the grey reef shark is an accurate predictor of more intense defensive behaviours (Fellows and Murchison 1967).

Displays evolve from non-communicative source behaviours ritualised via a five-step process shaped by natural selection (Tinbergen 1952).
Barlow (1974) proposed that the hunch display of the grey reef shark, like that of other predatory animals, was derived from movements associated with subjugating prey.

Accurate signalling of an animal’s motivation or state (‘‘display honesty’’) can be ensured by being costly, such as handicapping the signaller by diverting resources that would otherwise increase its fitness (Zahavi 1993).
However, accurate signalling must have benefits that exceed the costs for such behaviour patterns to be retained rather than eliminated through natural selection (Krebs and Davies 1993; Cuthill and Houston 1997).
Costs and benefits of agonistic displays in the grey reef and other sharks have not been analysed.

The hunch display is most often elicited in situations where food is not present, such as close approach to a shark by an underwater photographer not using bait as a lure, and probably plays a role in normal social encounters among grey reef sharks ( Johnson and Nelson 1973; Nelson and Johnson 1980).
This species is widespread throughout the tropical Indo-Pacific region and, despite its modest size (most individuals are smaller than 1.5m long), is avoided by or actively displaces most sympatric shark species, including solitary heterospecifics larger than itself ( Johnson 1978; Compagno 1984).
Myrberg (1991) suggested that distinctive fin markings in whaler and hammerhead sharks (Carcharhinidae and Sphyrnidae, respectively), especially on the relatively mobile pectoral and caudal fins, may serve as species recognition badges and/or accentuate signals, both social (intraspecific) and extrasocial (interspecific). The ventral apices of the grey reef shark’s pectoral fins are dusky and its caudal fin bears a distinct black rear margin (Last and Stevens 1994); in western Pacific and Indian Ocean representatives, the first dorsal fin features a conspicuous pale apex (Compagno 1984). Intriguingly, western Pacific and Indian Ocean grey reef sharks are more timid and less likely to display than those from French Polynesia or Micronesia (personal observation). Whether such regional differences in grey reef shark behaviour and markings are causally related is unknown.

The primary literature contains descriptions of agonistic displays similar in context and execution to that of the grey reef shark in several other shark species, including the bonnethead (Sphyrna tiburo), scalloped hammerhead (S. lewini), and blacktip shark (C. limbatus), but none as overt as that exhibited by the grey reef shark (Myrberg and Gruber 1974; Klimley 1985; Ritter and Godknecht 2000). Putative agonistic displays in several other shark species, including the Galapagos shark (C. galapagensis), Caribbean reef shark (C. perezi , white shark (Carcharodon carcharias), and shortfin mako (Isurus oxyrinchus), have been reported in the popular literature (Hobson 1964; Stoneman 1985; Compagno et al. 1989).

Confounding such anecdotal reports are recent studies showing that striped sharksuckers (
Echineis naucrates) can induce behaviours that resemble agonistic displays in the blacktip shark, bull shark (Carcharhinus leucas), Caribbean reef shark (Video clip 4), and possibly other shark species (Ritter 2002; Brunnschweiler 2006).
Blurring the picture further are dissimilar shark behaviours reported under competitive contexts, including tail slapping (whip-like lashing of the tail at or near the surface, hittingor directing splashes toward a receiver), which has been reported in sandtigersharks (Carcharias taurus) and white sharks, and breaching (leaping completely from the.water) and repetitive aerial gaping (RAG: rhythmic, exaggerated opening and closing of the jaws above the surface, apparently as a displacement activity), which have been reported in white sharks (Compagno et al. 1989; Klimley et al. 1996; Strong 1996).

The present work reviews relevant theoretical background issues and putative shark agonistic displays from the literature, augmented and extended with opportunistic observations by the author and collected from experienced, credible divers.

What constitutes an agonistic display is defined and the challenge of differentiating displays from non-displays in sharks is discussed. Previously undescribed agonistic displays or new display elements in six species of sharks are described and illustrated. All known shark agonistic displays are compared and general patterns are identified. Relative costs and benefits of shark agonistic displays in interactions with other sharks and divers are evaluated using payoff matrices. Shark–diver interaction scenarios in which sharks are most likely to exhibit agonistic displays are reviewed. A hierarchy of critical approach distances for sharks is hypothesised and diver responses to an agonistically displaying shark that may reduce likelihood of escalation to full attack are suggested. Recommendations for future work on shark agonistic behaviour are offered.


Shark measurements and display elements used in this review are:

Back arching – a sustained dorsal flexure of the back, so that it forms an inverted U, measured or estimated as the angle between the dorsal surface of the head and that of the pre-dorsal back (‘‘nape’’)
Body length (BL) – estimated total length, from snout tip to upper caudal fin apex
. Body shivering – a quivering or rapid (faster than 5 Hz) shaking of the entire body, usually combined with a lateral display
Body tilting or rolling – an exaggerated, sustained roll of the body around its longitudinal axis, up to plus/minus 45 degrees from vertical, either as a single action (tilt) or an oscillation (roll) (sensu Johnson and Nelson 1973; Klimley 1985; Ritter and Godknecht 2000)
. Charging – rapid (2 to 3þBL/s) approach toward a receiver, often terminated by veering away on a perpendicular course within a distance of 1–2BL
. Clasper flexion – rotation of a clasper (elasmobranch intromittant organ) anterolaterally, mimicking its erect state during copulation (sensu Myrberg and Gruber 1974)
Corkscrewing – rapid burst swimming into a tight looping trajectory while rotating the longitudinal body axis through nearly 180 degrees (sensu Klimley 1985)
. Flank displaying – a sustained perpendicular bodily orientation of signaller’s body toward a receiver, displaying its lateral surface
. Follow–give way – an agonistic interaction in which one animal follows another, which deviates from its course, implying social precedence of the following individual (sensu Myrberg and Gruber 1974)
. Gill pouch billowing – a sustained conspicuous, spinnaker-like expansion of the branchial region; equivalent to ‘‘gill-puff’’ of Myrberg and Gruber (1974)
. Give way – an agonistic interaction in which one animal deviates from its course at the approach of another, implying social precedence of the undeviating individual (sensu Myrberg and Gruber 1974)
Head shaking – rhythmic, exaggerated lateral shaking of the head, usually rapid (faster than 2 Hz) and through an arc of more than 30 degrees
. Jaw closing – a non-forceful bringing together the upper and lower jaws in an agonistic context (as compared to a bite, which is a forceful jaw closure during feeding)
Jaw gaping – a slow, exaggerated opening of the jaws (30–90 degrees, measured or estimated as the angle formed at the mouth commissure), conspicuously wider than during ram ventilation and in an agonistic context, resembling a yawn
Laterally exaggerated swimming – sustained , eel-like swimming motion
Looping – swimming in a spiral loop (sensu Johnson and Nelson 1973)
Open jawed tooth raking – a forceful strike with the upper jaw teeth causing injury to a receiver
. Palatoquadrate protrusion – forward extension of the upper jaw, causing it to separate from the cranium, and which may bare the upper dentition; categorized as present or not
Pectoral fin angle (PFA) – is measured or estimated from a horizontal plane perpendicular to the body axis at the level of the chord formed by a straight-line from pectoral fin origin to pectoral free rear tip (which is often easier to discern than the pectoral insertion in free-swimming sharks); equivalent to the ‘‘dihedral angle’’ of Wilga and Lauder (2000)
Pectoral fin depression – a sustained bilateral lowering of the pectoral fins from their usual position during swimming, quantified as pectoral fin angle
. Ramming with snout – a signaller using its rostrum to forcefully strike a receiver, often causing it to recoil or retreat; synonymous with ‘‘hit’’ of Myrberg and Gruber (1974)
. Rapid withdrawal – movement of a signaller away from a display receiver at _3BL/s.
. Rapid, tight pattern swimming – fast swimming marked by sudden and frequent changes in direction
. Reduced swimming efficiency – conspicuous reduction in forward locomotion, from almost normal swimming speed to nearly stationary in the water column, despite exaggerated swimming movements ( Johnson and Nelson 1973)
. Ritualistic jaw snapping – rhythmic (faster than 1 Hz), exaggerated opening and closing of the jaws, (30–90 degrees, measured or estimated as the angle formed at the mouth commissure)
. Snout elevation – a sustained raising of the snout from the resting position, measured or estimated as the angle between the dorsal surface of the rostrum and that of the epibranchial region (surface of the back above the gill slits) ( Johnson and Nelson 1973)
. Stand back – an agonistic interaction in which two oncoming animals deviate from their courses simultaneously, making it impossible to infer social precedence (sensu Compagno 2001)
Stiff or jerky movement – a conspicuous reduction or loss of fluidity of body and fin movements during locomotion, probably due to an increase in tonus of associated muscles
. Swimming speed – is measured in BL/s
Tail depression – a sustained vertical bending of the caudal peduncle, stiffly holding the causal fin downward, below the main body axis .
Tail flexure – a sustained lateral bending of the caudal peduncle, stiffly holding the causal fin to one side, away from the longitudinal axis ( Johnson and Nelson 1973)
Tail slapping – a whip-like lashing of the tail at or near the surface, hitting or directing splashes toward a competitor (Compagno et al. 1989; sensu Klimley et al. 1996)
. Torso thrusting – an exaggerated, sustained lateral flexure of the abdominal region into an S-shape, often accompanied by body tilting, sensu Klimley (1985)


Signal logistics and types

The term ‘‘agonism’’ was introduced by Scott and Fredericson (1951) to refer to any behaviour related to fighting and competition.
Agonism embraces a broad spectrum of behaviours related to animal conflict, including fighting, threat, defence, flight, freezing (remaining motionless to avoid detection), avoidance, and appeasement ( Johnson 1972; King 1973; Dewsbury 1978; McFarland 1987).

For all their variety, agonistic behaviours are united by occurrence during conflict situations (Scott 1971).

Conflict denotes a motivational state in which tendencies to perform more than one activity are expressed simultaneously (Manning 1979). Much agonistic behaviour is characterised by motivational conflict between attack and flight, often resulting in vacillation between approach and withdrawal (McFarland 1987).

A ‘‘display’’ is a stereotyped motor pattern involved in animal communication; displays are usually intraspecific and related species often have similar displays (McFarland 1987).

For well-armed contestants, physical combat bears significant risk of injury, which could compromise the feeding or reproductive abilities of both participants (Lorenz 1964). Therefore, in conflict situations, ritualised displays may be more adaptive than combat (Reichert 1982). Due to their conservative structure and the constraints of efficient swimming, sharks have few structures available for signalling. Their only suitable appendages are the relatively highly mobile pectoral and caudal fins (Myrberg 1991) and, in males, the claspers (Myrberg and Gruber 1974; Klimley 1985). Signals are ritualised from non-communicative source behaviours via a process of simplification, exaggeration, decreased performance speed, repetition, and increased stereotypy (Tinbergen 1952). Thus, by definition, shark agonistic displays are conspicuous MAPs elicited in conflict situations, usually between or among conspecifics.

The hunch display of grey reef sharks, however, is not known to be performed toward conspecifics, but is known to be directed toward humans ( Johnson and Nelson 1973; Nelson et al. 1986), great hammerhead sharks (Sphyrna mokarran) and moray eels ( Johnson 1978).
Similarly, blacktip sharks have been documented displaying toward scuba divers before accelerating away (Ritter and Godknecht 2000; Stafford-Deitsch 2000). Since shark agonistic displays are unlikely to have evolved in response to diver encroachment (within the last 50 years), it is likely that shark species perform agonistic displays the same way toward natural predators or competitors as toward humans. This performance conservatism is useful, as it enables scientific study of shark agonistic displays.

Recognising shark agonistic displays

Complex biological structures, such as fins, are often multifunctional (Bock and von Wahlert 1998).
Caudal and pectoral fins of sharks play important and active roles in shark locomotion (Bone 1988).
This raises methodological questions about differentiating between use of these fins for locomotion versus signalling.

Use of pectoral fins by sharks for turning is conspicuously different from their use during agonistic displays
(Barlow 1974).
Qualitatively, during turning in free-swimming sharks in the wild, the pectoral fins are flexed slowly in a fluid, short-duration (usually shorter than 1 s) motion and PFA of the left and right pectoral fins are unequal, with the fin on the side to which a shark turns usually depressed more strongly than the other, although this is not always the case in slow-moving sharks.
This biased pectoral fin depression is contrary to Figure 11 in Wilga and Lauder (2000), which may reflect simplification for clarity or an artefact of inhibited steering in the confines of a laboratory tank. In contrast, during shark agonistic displays, both pectoral fins are held stiffly downward for prolonged periods (usually 5 to 120þs), and bilaterally. Such pectoral fin depression, probably derived from postures assumed by a shark bracing its body in the water column during prey dismemberment would be expected to increase yaw stability and decrease agility. …….

However, sustained bilateral pectoral fin depression, performed in a conflict situation, and accompanied by other common shark agonistic display elements (described subsequently) greatly increase the likelihood of correctly identifying a shark agonistic display.

Species in which displays are known

Agonistic displays are known in at least 23 species of sharks, representing six families.

Sandtiger shark.
This species exhibits a variety of agonistic behaviours not reported by Compagno et al. (1989) in western Indian Ocean individuals. Among the most obvious of these in western North Atlantic individuals is pectoral fin depression accompanied by flank displaying. As in grey reef sharks, pectoral fin depression is elicited in approach–withdrawal conflict situations and is a graded response, with intensity of PFA increasing from 30 degrees to 50 degrees with diver proximity and degree of escape route restriction.
This species has attacked divers off South Africa and elsewhere, delivering minimalist bites (not removing flesh) similar to those inflicted on captive conspecifics (Compagno 2001). Another putative agonistic behaviour in this species, often performed in conjunction with pectoral fin depression, is a loud shotgun-like report, generated by exaggerated tail beats during rapid withdrawal that is termed ‘‘tail popping’’ by North Carolina wreck divers (R. Allen, personal communication). Preliminary field tests off North Carolina suggest that likelihood of tail popping increases with speed, directness, and angle of diver approach. Rapid frontal approaches at a downward angle of 30 degrees from the animal’s longitudinal axis seemed to produce the strongest tail popping response. Similar sounds generated by the caudal fin of this species have also been reported from Australia, where groups of sandtigers were observed apparently herding a school of carangids (Cliff 1988). However, it is probable that sound production in this species is caused, at least in part, by cavitation in the wake of its large subterminal lobe), which may make it an artefact of rapid acceleration rather than a display element. Systematic testing of tail popping under agonistic and non-agonistic contexts is needed to resolve this matter. Such testing could include speaker playback of tail popping, which would allow separation of visual and other shark-related cues from the potential stimulus-specific role of popping, analogous to birdsong playback to distinguish mate-attraction from competitor-repulsion functions (e.g. Searcy and Nowicki 2000; Vehrencamp 2000).

White shark.

A variety of named and described agonistic behaviours have been reported in this species, but without supporting evidence (Compagno 2001; Martin et al. 2005).
A photo of a white shark hunch display featuring strong pectoral fin depression (PFA¼_70_), was taken off Dyer Island, South Africa in July 2003 (Figure 3i). A possible hunch display by a white shark, apparently contesting access to speared fish and followed by an attack on the spearfisher, was reported in South Africa by Jones (1991), while Collier (2003) identifies four cases in which this species appears to have attacked humans off California in agonistic contexts. Between July 2002 and July 2005, the following agonistic behaviours were performed by white sharks toward conspecifics at Seal Island in False Bay, South Africa: 22 give ways, four follow-give ways, three hunch-give ways, and two stand backs. As these behaviours are well defined and illustrated elsewhere (Myrberg and Gruber 1974; Martin 2003; Martin and Martin 2006), they will not be further described here.. On 27 June 2005, after having been thwarted by the bait handler in repeated attempts to consume the bait, a 3.5m male white shark twice performed RAG displays completely consistent in form and context with those described and illustrated in Strong (1996).

Silvertip shark.
Encroachment by underwater photographers and pursuit by a submersible elicited a partial agonistic display in this species. This display featured slightly stiff-bodied swimming and brief pectoral fin depression (Eibl-Eibesfeldt and Hass 1959, Nelson et al. 1986). Diver pursuit elicited a more elaborate, highly stereotyped agonistic display in at least five different individuals (1.4–1.8mBL) in eight out of 28 (28.6%) trials on non-feeding animals at several offshore reefs on the Great Barrier Reef, Australia. In each successful trial, when a distance of about 2BL was broached by a diver, the test subject accelerated away to a distance of about 15 m. Each shark then turned sharply and charged toward the observer at 3BL/s. When the shark reached a distance of about 2 BL, it strongly depressed its pectoral fins, turned broadside, vertically depressed its caudal peduncle, and —while holding this posture and moving at greatly reduced speed (slower than 0.5BL/s)— gaped its jaws rhythmically (at approx.1 Hz) and vibrated its entire body in a shiver-like motion. This shiver accentuated the white fin markings characteristic of the species and was maintained until the posterior margin of the animal’s pectoral fin base passed the observer’s mid-line. The shark then ceased shivering, raised its caudal fin and accelerated away at 3BL/s with both its pectoral fins strongly depressed . In one trial, a 1.7m female silvertip shark attacked before fleeing; the observer was uninjured but the nose-cone of the DPV was raked by the shark’s upper teeth. Repeated attempts to approach this species for photographic purposes in the Chagos Archipelago, central Indian Ocean, elicited essentially the same display Emery, personal communication).

Galapagos shark.
Aspects of an agonistic display, featuring laterally exaggerated swimming, have been reported in this species (Hobson 1964). During this display, the shark’s head swung through an arc of >90 degrees, which Hobson (1964) interpreted as an attempt by the animal to increase its visual field to maintain visual contact with the diver following it. Sustained diver pursuit elicited a more extreme version of this display on four occasions in non-feeding subadult Galapagos sharks (1.5–2mTL) at Lord Howe Island, Australia. This display was elicited when the observer repeatedly approached from behind to within 2BL of the shark. It featured laterally exaggerated, sub-anguilliform swimming movements, strong bilateral pectoral fin depression (PFA¼ 40 to 60_), gill pouch billowing, and exaggerated jaw gaping at about 0.5 Hz .The display was terminated when all postural and locomotory behavioural elements ceased, except bilateral pectoral fin depression, which was sustained during rapid withdrawal at _3BL/s. A partial agonistic display by a juvenile Galapagos shark, featuring flank displaying and strong pectoral fin depression (PFA¼_70_) was photographed in October 2003 at Bassas da India Atoll, Mozambique Channel (Figure 3l), which may be a nursery area for this species (Hammerschlag and Fallows 2005).

Caribbean reef shark.
Stoneman (1985) reported more than 450 observations of agonistic behaviour, some of which (fraction cannot be determined, since full data are not available) resulted in attacks, in what he identifies as Carcharhinus limbatus (blacktip sharks). However, Stoneman seems to consistently mistake Caribbean reef sharks for this species (Meltzoff 1988; Gruber 1993). Unfortunately, Stoneman has not published his observations, so they are unavailable for independent analysis. Consultation with Stoneman’s scientific advisor indicated that the displays they attributed to C. limbatus featured brief, 1 to 1.2-second pectoral fin dips, punctuated by frequent changes in direction (A. Emery, personal communication). Diver-acclimatised Caribbean reef sharks in the Bahamas were observed in the presence and absence of bait by the author using scuba. Despite high concentrations of sharks (approx. 60 animals in an area smaller than 2500m2), no clear agonistic behaviour was observed without baits, but repeated brief pectoral fin dips (PFA= approx -15 degrees, period=1 Hz) followed a second or so later by 90-degree changes in direction were evident when bait was present, supporting the proposition that the behaviours attributed by Stoneman (1985) to C. limbatus were performed by Caribbean reef sharks. Scharfer (2003) reported similar agonistic behaviour patterns in Caribbean reef sharks at another long-term shark feeding site in the Bahamas, noting that frequency of agonistic acts by sharks depended on feeding technique and that female sharks bumped camera ports (a possible agonistic behaviour) more frequently than males. Systematic observations, preferably of less diver-acclimatised sharks, are needed to verify agonistic behaviour patterns in Caribbean reef sharks.

Sandbar shark.
Agonistic displays have not previously been reported in this species, which is a congener of the grey reef shark and ten other species from which such displays have been reported. Partial agonistic displays by subadult sandbar sharks (1.2–1.5mBL), in mixed aggregation with Galapagos sharks (1.8–2mBL), and in the presence of baits, were observed from a shark cage off the north shore of Oahu, Hawaii. This display appeared to be directed toward the observer/cage and featured partial flank display combined with moderately strong pectoral fin depression, increasing from an initial PFA of 5 degrees to a final PFA of 60_ degrees. In more intense versions of this display in Hawaiian waters, subadult sandbar sharks charged toward a diver in the presence of bait at 2–3BL/s. During such charges, the pectoral fins were strongly depressed and, at a distance of 1.5 BL, males veered away from the diver on a perpendicular course, while females continued toward the diver and sometimes rammed him with the snout tip before withdrawing rapidly (D. Fleetham, personal communication).

None of the preceding six species exhibited palatoquadrate protrusion as a component of agonistic displays.


Sensory channels of communication

Sharks are endowed with a battery of well-developed senses, responsive to mechanical, chemical, visual, and electrical stimuli (see Hueter et al. 2004, for a recent review).
In theory, any or all of these sensory channels are available for communication (Dusenbery 1992). However, the marine operating environment imposes signal attenuation and limits signal functional range, ability to go around obstacles, transmission speed, and persistence as well as receivers’ ease of locating signal source (Bradbury and Vehrencamp 1998).

Visual signals propagate at the speed of light within limits set by the optical properties of the transmission medium, are directional and non-persistent (McGregor 2005).
Sharks possess a high order of visual sensitivity and acuity, with pupillary dilation and tapetal plates affording some ten times greater photic sensitivity than human eyes under scotopic conditions and at least some species exhibit a well-developed Purkinje shift and true colour vision (Gruber and Cohen 1978; Hueter et al. 2004). There is no reason to believe that sharks cannot see to the limit of underwater visibility which, depending upon habitat, may be as far as 30m (Hueter and Gilbert 1991).


Thus, pending further data, it is provisionally suggested that
vision may be the primary sensory channel employed by sharks for agonistic signalling.

What information is conveyed?

Agonistic encounters are framed by uncertainties between or among the combatants.

Agonistic displays provide information about the likely intent and level of commitment of the signaller and may provide information on relative fighting ability.

For example, agonistic displays may emphasise body size, weapons, or status badges, and may also include a tactical element, placing the signaller in a better position to attack or flee (Bradbury and Vehrencamp 1998).
Agonistic displays are stabilised by natural selection if the costs are borne by the signaller, in what are termed ‘‘vulnerability handicaps’’ (Vehrencamp 2000). Size is a good indicator of fighting ability, such that competitors of grossly dissimilar size rarely need to fight or display (Krebs and Davies 1993). If, however, combatants are similar in size, ritualised agonistic displays are likely to occur (Bradbury and Vehrencamp 1998).

In assessing the validity of an agonistic display, two sources of uncertainty arise, (1) receiver estimation of signal magnitude, and (2) degree to which a display indicates the likely outcome of combat (Bradbury and Vehrencamp 1998).
Agonistic displays performed at close range to the receiver are likely to be more effective, but they also bear greater risk (Zahavi 1977). Intention movements are incomplete initial phases of behaviour patterns and many threat displays, such as jaw gapes, are ritualized from attack or defensive behaviours (McFarland 1987; Hauser and Nelson 1991). Frontal and lateral agonistic displays occur as a result of motivational conflict between attacking and fleeing, with frontal displays indicative of a tendency toward attack and lateral displays of a tendency toward escape (Fine et al. 1977). Broadside or lateral agonistic displays are widespread among vertebrates (Thomas et al. 1965; Carpenter 1978; Hinde 1982), probably because they emphasise size by presenting the largest possible aspect; such displays also place the signaller at elevated risk of attack, which could signal the sender’s confidence in its fighting ability. Complicating these issues further, individual combatants may recognise a degree of relatedness or have prior experience of one another, but researchers are rarely privy to such parameters (especially in wild or free-ranging subjects), introducing much uncertainty into interpreting isolated agonistic interactions among study subjects


Clear agonistic displays are difficult to elicit in sharks, because they require sustained (several minutes) oriented pursuit.
Drawing on visual and possibly other cues, sharks in the wild use their superior swimming ability to set the preferred minimum distance between themselves and divers; if a diver attempts to approach, most simply swim away to avoid contact. However, in some cases, a shark that is approached by a diver will not retreat and will attack instead.

Traditionally, it was thought that shark attacks on humans were motivated by hunger.
Baldridge and Williams (1969) were the first to suggest that the character of many shark attacks seemed more consistent with fighting than feeding. Analyses of 1165 case histories archived in the International Shark Attack File concluded that 50–75% of shark attacks on humans were motivated by something other than a desire to feed (Baldridge 1974b, 1988).

The hunch display of the grey reef shark appears defensive in nature and often precedes a rapid, slashing attack with little or no removal of flesh (Fellows and Murchison 1967; Johnson and Nelson 1973; Nelson et al. 1986).
Agonistic displays by sharks have preceded attacks on humans in at least five other species: sandtiger (Compagno et al. 1989; Compagno 2001; and present study), white shark (Jones 1991; Collier 2003), silvertip (Nelson et al. 1986; and present study), Caribbean reef (Stoneman 1985), and smooth hammerhead (Sphyrna zygaena) (Miller and Collier 1981; Collier 2003).

Yet the majority of agonistic displays do not result in an all-out attack.

Nelson (1981, 1983) suggested that sharks may sometimes attack humans because they are regarded as competition for food or another resource.
Shark agonistic displays are rarely elicited during feeding or in the presence of bait, which suggests that access to a food resource is not usually a precipitating factor and that divers are not regarded by sharks as competitors. However, in two attacks on divers involving white sharks and one involving a smooth hammerhead, agonistically displaying sharks appeared to contest access to speared fish ( Jones 1991; Collier 2003).

One often-raised possibility is that a displaying shark is defending a geographic area and its resources (territory) for its exclusive use (Baldridge and Williams 1969; Johnson and Nelson 1973; Nelson 1981, 1983).

Defining territoriality in a way that all ethologists agree upon is difficult because different workers emphasise different attributes to different degrees, depending upon their specific research questions. For purposes of the present review, territoriality will be defined following Wittenberger (1981) as more or less exclusive control of a geographic area by an individual or group of individuals by means of implied or overt aggression.

There is no record of reef-dwelling sharks agonistically displaying toward conspecifics — which, having most similar resource needs and exploitation capabilities, constitute the most intense ecological competitors (Nelson et al. 1986).
Direct, rapid approaches by divers are most likely to elicit shark agonistic displays, especially if potential escape routes are restricted or absent.
This observation is consistent with the hypothesis (Fellows and Murchison 1967; Johnson and Nelson 1973; Nelson et al. 1986) that these displays signal defensive threat. The open-mouthed, slashing attack that sometimes follows persistent movement toward a displaying grey reef shark, as well as the rapid withdrawal that terminates the agonistic display, are likewise consistent with self-defence rather than feeding.

Pending further evidence on the issue of site-dependent aggression, it seems likely that, i
n most cases, an agonistically displaying shark is motivated by defence of self rather than of a food resource or territory.

Common features of shark agonistic displays

Agonistic displays of a few shark species exhibit unique elements (e.g. body shiver in the silvertip or corkscrewing in the scalloped hammerhead, S. lewini ), but a number of elements appear common to the agonistic displays of most .
The seven most common features (with percentage occurrence in parentheses) include:
(1) stiff or jerky movements (100%),

(2) sustained bilateral pectoral fin depression (100%),
(3) rapid withdrawal (83%),
(4) jaw gaping (70%),
(5) reduced swimming efficiency (52%),
(6) flank displaying (48%),

(7) laterally exaggerated swimming movements (13%).

All these elements are found in the hunch display of the grey reef shark, indicating that its highly overt agonistic display represents an extreme along a continuum of agonistic behaviours in sharks, which range from subtle (in the Caribbean reef shark) to overt (in the grey reef, silvertip, and Galapagos sharks).

Intriguingly, agonistic behaviours that resemble those of sharks in form (arched back, lowered flippers, ritualised jaw gaping, and flank displaying) and context have been reported in at least three species of cetaceans, the spinner dolphin (
Stenella attenuata), the bottlenose dolphin (Tursiops truncatus), and the humpback whale (Megaptera novaeangliae) (Overstrom 1983; Connor and Peterson 1994; Johnson and Norris 1994).

Agonistic displays of many unrelated species exhibit convergent properties that honestly inform rivals about the size and fighting capacity of the signaller (Parker 1974).
Whether these cetacean agonistic display elements represent behavioural mimicry of sharks, shared hydrodynamic and morphological constraints, or some other form of convergence is not clear. But in the agonistic displays of both sharks and cetaceans, presentation of offensive weapons and the largest possible aspect come at the cost of decreased swimming ability.

Shark–human interactions

Human pursuit and proximity are sign stimuli known to elicit agonistic displays in sharks.
However, intensity of agonistic responses by sharks is context dependent. Johnson and Nelson (1973) found the likelihood that a grey reef shark performs an agonistic display increased with speed and directness of diver approach as well as with degree of escape route restriction. This pattern appears to hold true for most species of actively swimming (i.e. non-benthic) sharks (personal observation).

Almost any object in the environment – bottom features, divers, other sharks, or possibly a boat – can restrict a shark’s escape route and may render it more likely to perform an agonistic display (personal observation).

For example, while swimming near the bottom, individual sharks of
most species appear agitated by any diver that maintains position a few metres above the animal’s back, a situation in which diver proximity and pursuit, a suboptimal view of the diver, and restriction of escape routes by the bottom combine to increase the animal’s stress and the likelihood to perform an agonistic display (personal observation; C. Harvey-Clark, personal communication). A tightly packed group of scuba divers more readily provokes agonistic display in solitary sharks than does a single diver or dispersed group of divers and thus may be considered a supernormal stimulus (Bolhuis and Giraldeau 2005).

Use of bait is well known to induce sharks to approach divers more closely than they do without such provisioning (Nelson 1977; Stafford-Deitsch 2000), although use of bait to stage ‘‘feedings’’ for diving tourists is highly controversial (Brylske 2000; Carwardine 2004).

Single-species groups of sharks that have become acclimatised to being fed individually by taking bait impaled on the end of a feeding pole appear relaxed and—despite close proximity of divers and conspecifics—rarely, if ever, display agonistically. In contrast, shark agonistic displays may be elicited toward divers during mixed species feeding events, especially when food is thrown or scattered into the water. Under such conditions, proximity of the bottom and nearby sharks may restrict escape route of an agonistically displaying shark and increase risk of attack against a nearby diver.

Interpreting the significance of these differences is hampered by general lack of data on the extent of learning from past inter- or intraspecific encounters among individual sharks, including the effects of previous agonistic wins and losses.
However, as with any wildlife encounter, ability to anticipate and avoid shark–human encounter scenarios known to agitate sharks, such as those described here, can help reduce the inherent risks of such encounters.
Shark–diver interactions are inherently risky but the level of danger can be limited by a diver’s ability to recognise signs that a shark is becoming agitated and respond appropriately (personal observation).

Under unbaited conditions, most sharks simply avoid divers.

However, some individuals, particularly of relatively large and actively predatory species, appear curious about divers and may approach closely.
It is probably safest for a diver to allow a shark to approach on its own terms; attempts to approach it will likely cause the animal to flee. A mildly agitated shark may accelerate away for a few body lengths and then resume a more relaxed cruising pace. As a shark becomes moderately agitated by diver encroachment, it often changes course frequently, using short, jerky movements. With increasing agitation, common elements of shark agonistic displays begin to appear.

Most conspicuous of these are rapid, jerky movements and sustained, bilateral pectoral fin depression.
The more agitated the shark, the more stiff and jerky its swimming movements and the more strongly its pectoral fins are depressed. These signs typically escalate in the regular – and thus predictable – patterns outlined here. Intensity of shark agitation increases with diver proximity, speed and directness of diver approach, and the number of sharks. However, in the vast majority of encounters, shark agonistic displays and signs of shark agitation reduce in frequency and intensity with diver withdrawal.

Thus, to a certain extent, an experienced diver can mediate the intensity of shark encounters through his own actions.
It is, however, absolutely vital that divers maintain visual contact with all sharks in the immediate vicinity throughout an encounter. Should a diver be unable to see all sharks in the vicinity at all times, prudence dictates leaving the water as quietly and efficiently as possible.
Based on results presented here, it seems likely that a shark performing an agonistic display is signalling that it is in a highly agitated state and on the verge of either attacking or fleeing; if the receiver persists, the shark will attack rather than flee. Such ritualized threats appear to be mostly bluff (the shark usually flees without attacking), but prudence dictates that one should not depend on this. This comparative behavioural review identifies common elements in shark agonistic displays that could assist divers and other ocean users to recognise whether a shark is becoming agitated and may inform human responses to an agonistically displaying shark.

Critical approach distances and diver safety

The best defence, when faced with a shark that appears to be agitated is to avoid exacerbating the situation. If a shark appears agitated (exhibits stiff, jerky movements and/or frequent direction changes) but has not yet begun to display, maintaining visual contact with the animal and withdrawing are likely to reduce signs of the animal’s agitation and stimulate it to withdraw from the area (personal observation).

For divers faced with an agonistically displaying shark, the outcomes of the present review and model suggest that the following safety recommendations may be of assistance:
(1) Face the displaying shark and maintain eye contact

(2) Do not touch, approach, or photograph a displaying shark
(3) If carrying speared fish or collected marine animals, surrender them immediately

(4) Back away from the displaying shark, using smooth purposeful movements; the shark will most likely seize the opportunity to withdraw
(5) Remain on the bottom, with your back toward a reef, wreck, or – if nothing else is available – your dive buddy

(6) Return to, or immediately below (if boat diving), your pre-planned exit point and exit the water unhurriedly but efficiently
(7) During surfacing in open water and while at the surface, check water around and beneath in case the shark follows

(8) If the shark follows, be ready to fend it off with any available object; avoid using your bare hands, if possible, as the shark’s rough skin and teeth may cause injury.

It is clear that much remains to be learned about shark agonistic displays and also about their perception of and interactions with humans.
Further study of agonistic behaviour in sharks is likely to pay dividends not only in a richer understanding of shark social dynamics but also in increased safety in shark–human interactions. With knowledge and care, sharks can be approached in their natural habitat and observed at close range in relative safety and with minimal observer impact. Such direct observations are likely to provide new insights into subtler aspects of shark social behaviour.

Are you still with me?
Here’s what I learn from reading this.
  • Remember: it seems likely that, in most cases, an agonistically displaying shark is motivated by defense of self rather than of a food resource or territory. Thus, rule number one is to avoid any behavior that a Shark may interpret as being threatening.
  • Learn to interpret Shark behavior by being alert to any of the tell-tale displays described above: Sharks are highly transparent and predictable if you know what to look for – and that includes accepting that predatory Sharks are predatory Sharks, and not some romantic figment of your imagination! Of interest, on top of the species-specific behavior described above, the paper contains a table describing the following agonistic display elements for Tigers: stiff or jerky movements; pectoral fin depression; jaw gaping; flank displaying; reduced swimming efficiency; ramming with snout and rapid withdrawal – and for Bulls: pectoral fin depression; snout elevation and jaw gaping, to which I would add head shaking; jaw closing and above all, ramming with snout (!). If in doubt, back off!
  • Avoid getting too close to the animals. Apart from the painfully trivial and yet apparently often forgotten simple fact that in order for there to be an incident, a human and a Shark need to be in the same place at the same time, Sharks have a personal space which in scientific terms is called idiosphere, defined as the volume immediately surrounding an individual animal (of any species) that it will defend if broached without appropriate signalling or permission. Learn to identify that space by observing the Shark’s reaction and retreat if you have breached its confines. Also, since that space is likely to vary with the degree at which an individual Shark is habituated to divers, always let the Shark swim to you (keep eye contact) by remaining stationary.
  • Related but not identical: avoid crowding the animals. What a Shark may interpret as being an obstacle impeding its withdrawal may not be readily discernible to you, as in another group of divers or, as exemplified by the paper, swimming above it whilst it is close to the bottom (Tiger Beach anybody?). By the same token, do not assertively swim towards, let alone chase any Shark, ever.
  • Do not touch. Sharks are not pets, reefs are not petting zoos and Sharks are never, ever, underwater scooters!
  • Selected baiting is better than indiscriminate dumping. This is totally in line with what we do in Fiji, as we always keep control of the food and can discontinue feeding as soon as we sense (and we sometimes do) that the situation is becoming tense.
  • Avoid creating competitive feeding situations. Remember: There is no record of reef-dwelling sharks agonistically displaying toward conspecifics, but still, mistakes do happen and excited Sharks may bite each other, and you, by accident. Moreover, a Shark may regard you as a competitor and act aggressively assertively!
  • If in trouble, push back. Our rule for image hunters: as long as you keep your camera between the mouth of the Shark and yourself, you're OK!
Easy – right?
Which once again begs the question, what are those operators thinking who facilitate those unsupervised interactive encounters with large predatory sharks?
Not much it appears - it's always the same combo: ignorance and greed!

Let’s hope that this post will contribute to eliminating the former!

No comments: